THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No. 7, Issue of March 5, pp. 4437-4441, 1991 Printed in U.S.A.

Self-Association of the Plasma Membrane-associated Clathrin Assembly Protein AP-2”

(Received for publication, July 9, 1 990)

Kenneth A. Beck and James H. Keen$ From the Fels Institute for Cancer Research and Molecular Biology and the Department of Biochemistry, Temple University School of Medicine, Philadelphia, Pennsylvania 19140

A ~ l f - ~ ~ i a t i o n reaction involving the plasma membrane-associated clathrin assembly protein AP-2 has been detected by incubating AP-2 alone under solution conditions that would favor the assembly of complete coat structures if clathrin were present. Self- association was rapid, unaffected by nonionic deter- gents, readily reversible, and gave rise to sedimentable aggregates. Only the AP subtype AP-2 exhibited self- association: the structurally or functionally related as- sembly proteins AP- 1 and AP-3 and unrelated proteins neither self-associated nor were incorporated into the AP-2 aggregate. AP-2 interactions responsible for self-association were of high affinity, with an apparent K d of approximately lo-’ M. By proteolytic dissection, the self-association domain was localized to the core of the molecule containing the intact 50- and 16-kDa polypeptides in association with the truncated 60-66- kDa moieties of the parent a/@ polypeptides. Self-as- sociation of the intact AP-2 molecule was pH-depend- ent, exhibiting an apparent pKa = 7.4. While it is unlikely that the large AP-2 aggregates formed in solution are themselves biologically relevant struc- tures, the AP-2 interactions involved in their forma- tion have properties consistent with their occurrence in intact cells and thus may be important in cellular functions of the plasma membrane-localized assembly protein.

Clathrin-coated pits and vesicles are ubiquitous cellular structures that are involved in both the endocytosis of various extracellular macromolecules and in Golgi sorting and trans- port (14) . The clathrin coats of these in structures contain two major components. The first of these is clathrin which is found in its unassembled (dissociated) form as a three-legged triskelion-shaped structure (5 , 6). These clathrin triskelions represent the major structural units of the coat and are responsible for its characteristic polygonal lattice morphology (3,5). The second component of the coat is a group of proteins referred to as clathrin assembly proteins (AP)’ by virtue of their ability to promote the assembly of clathrin into coat

* This work was supported by National Institutes of Health Grant GM-28526 (to J. H. K,), The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Fels Institute for Cancer Research and Molecular Biology and the Dept. of Biochem- istry, 3420 N. Broad St., Temple University School of Medicine, Philadelphia, PA 19140.

The abbreviations used are: AP, clathrin assembly protein; MES, 2-[N-morpholino]ethanesulfonic acid; HEPES, 4-(2-hydroxyethyl)- 1-piperazin~thanesulfonic acid.

structures in vitro (7,8). In bovine brain three such complexes of AP, designated AP-1, AP-2, and AP-180 (9-12), have been identified. Immunofluorescence localization studies (13, 14) have indicated that the AP-2 subtype is associated with coated pits of the plasma membrane whereas AP-1 is present in the Golgi apparatus.

Clathrin coats are peripherally associated with the cyto- plasmic surfaces of membranes and can be stripped from isolated coated vesicles under a number of conditions (15-17) including 2 M urea, high concentrations of protonated amines at neutral pH (0.5 M Tris-HCI, pH 7.0), and high pH buffers (usually 10 mM Tris-HC1, pH 8.5). These treatments disrupt the interactions of the coat proteins with both the membrane and with each other and therefore give rise to in~vidual free clathrin triskelions and AP molecules. They are also nonde- naturing and hence produce free coat components that are structurally and functionally intact. Mixtures of purified clathrin and AP in 0.5 M Tris-HC1 will assemble into spherical coat structures resembling the coats found on coated vesicles in their polygonal lattice morpholo~ when the Tris-HCl is removed by dialysis (7,B).

In this study we show that when AP is treated alone under such assembly conditions, a self-association reaction is re- vealed. Further investigation demonstrates that this reaction is unique to the AP-2 subtype and has characteristics consist- ent with its occurrence in intact cells. The nature and prop- erties of this self-association reaction are described and its potential relevance to AP-2 functions in cells is discussed.

EXPERIMENTAL PROCEDURES

Materials-AP was isolated by Superose 6B gel filtration as de- scribed previously (7, ll). In most of the experiments, material corresponding to fractions 36-38 in Fig. 2 of Ref. 11 were used. For experiments with pure AP-2 and partially purified AP-1, fractions corresponding to 38-40 in Fig. 2 of Ref. 11 were pooled and fraction- ated by clathrin-Sepharose chromatography as described elsewhere (11). For experiments involving AP 180, fractions corresponding to 35 and 36 in Fig. 2 of Ref. 11 were used. Trypsin (~-1-tosylamido-2- phenylethyl chloromethyl ketone-treated, 241 units/mg) was from Worth in~on (Freehold, NJ).

Methods-AP self-association experiments were carried out either by dialysis of AP (in 0.5 M Tris-HCl) against 0.1 M sodium MES, pH 6.5 (buffer A) or by diluting concentrated samples of AP (in 10 mM Tris-HCI, pH 8.5) with buffer A. Dialysis experiments were carried out by placing samples of AP (100 pg/ml in 0.7 ml of 0.5 M Tris-HC1, pH 7.0) in collodion dialysis bags (UH100/25, Schleicher and Schuell) and dialyzing against 300 ml of buffer A at 4 “C. For dilution experi- ments AP was first precipitated with 50% saturated ammonium sulfate, resuspended in a small volume of 10 mM Tris-HCl, pH 8.5 (0.3 ml), and dialyzed against this buffer overnight to remove residual ammonium sulfate. Concentrated samples of AP prepared in this way (typically 1-2 mg/ml) showed little evidence of aggregation (2% of what is seen in buffer A). Aggregation experiments were carried out by diluting these samples with buffer A (at 4 “C) such that the final AP concentration was typically 100 pg/ml and the Tris-HCI concen-

4437

by guest on April 2, 2018

http://ww

w.jbc.org/

Dow

nloaded from

by guest on April 2, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Self-Association of Assembly Protein AP-2 4439

0 10 20 30 Tlms (rnl“)

S P S P S P 2 3 4 8

-AP 16

I2O ‘ , I C c

ITrla-HCII (mu)

FIG. 2. Quantitation of AP self-association. Concentrated samples of A P (1.84 mg/ml) in 10 mM Tris-HCI, pH 8.5, were diluted 15-fold with buffer A containing various concentrations of Tris-HC1, pH 6.5. A, sample absorbance was measured at various times after dilution. The final Tris-HCI concentrations were 0.3 (O), 25.3 (O), 50.3 (A), and 100.3 mM (0). The zero time value reflects dilution into an equivalent volume of 10 mM Tris-HC1, pH 8.5. B, the amount of self-associated A P in the samples was quantitated by centrifugation (see “Methods”). The lanes correspond to supernatant (s) and pellet ( p ) fractions of samples treated in the presence of 0.3 (lane I ) , 25.3 (lane 2), 50.3 (lane 3), and 100.3 mM (lane 4 ) Tris-HCI. The positions of the AP-2 subunits are indicated (right). C, the percentage of sedimented A P (both the 100-kDa (0) and 50-kDa (0) subunits) in the samples shown in B, was quantitated by densitometry and is plotted as a function of the Tris-HC1 concentration. Also shown is the amount of self-associated A P as determined by Asso (Fig. lA), relative to the control maintained in 0.3 mM Tris-HCI (A).

FIG. 3. AP self-association as a function of AP concentra- tion. Samples of A P were diluted with buffer A to the indicated final concentrations and their A350 was measured after 30 min. The Asso of unaggregated protein, assessed by dilution into an equivalent volume of 10 mM Tris, pH 8.5, was negligible (<5%). The inset is an expansion of lowest concentration range.

sharp critical concentration cannot be rigorously concluded because of potential limitations in the assay. However, the apparent lower limit of complete self-association at 25 pg/ml or ~ 7 5 nM corresponds to a dissociation constant of approx- imately lo-’ M. This is supported by the extrapolated inter- cept on the abscissa which yields a value of 6 pg/ml or 17.5

nM for the equilibrium concentration of free monomer. AP self-association was maximal at pH 6.5 and decreased

sharply with increasing pH (Fig. 4). The effect of pH on the AP concentration dependence of self-association (Fig. 4, in- set) is to alter the slope and is therefore consistent with the titration of an active species rather than a pH-dependent reduction in the affinity of AP-AP interactions or increase in the critical concentration (34). Hence, the steepness of the curve in Fig. 4, which agrees well with a titration curve calculated for a single species (Fig. 4, dashed l i n e ) , is consist- ent with the existence of a discrete site on AP-2 that is required for the reaction; presumably other interacting sites that are involved in the aggregation reaction are also present on the AP-2 molecule. The pK, derived from the titration curve, 7.4, indicates that self-association is most sensitive within the range in which cytoplasmic pH is known to vary

To determine if the AP-AP interactions leading to self- association are reversible we first prepared AP aggregates by dilution with 0.1 M sodium MES solutions at either pH 6.5 or 7.5 and measured their formation by A350 (Fig. 5). The aggre- gated protein in the pH 6.5 sample was separated by centrif- ugation and resuspended in 10 mM Tris-HC1, pH 8.5. The low A350 of the resulting solution (details in figure legend) indi- cated that the AP had indeed dissociated. When this sample was again induced to reassociate at either pH 6.5 or 7.5, the extent of self-association was quantitatively equivalent to that observed in the first round of aggregation (Fig. 5).

A domain structure of AP-2 has been elucidated by a combination of proteolytic studies and electron microscopy. Deep-etch electron microscopy of pure AP-2 reveals a tripar- tite structure consisting of a large central domain with two smaller appendages (21). Treatment of AP-2 with elastase or trypsin results in the generation of two separable products previously referred to (23) as heavy and light mero-AP. The larger of these two fragments, heavy mero-AP, corresponds to the large central domain of the native protein and is a complex of the intact 50- and 16-kDa subunits of AP-2 as well as the 60-66-kDa proteolytic fragments of the parent 100-115-kDa subunits (22, 23). The term light mero-AP de- scribes the 30-36-kDa framnents derived from the 100-kDa

(pH 7.0-7.4 (29)).

- subunits, and corresponds to the smaller appendages.

I I I I

To

0 ‘ I I I 6 7 8 9

PH FIG. 4. Effect of pH on AP self-association. A P self-associa-

tion was induced by diluting concentrated samples of A P with 0.1 M sodium MES of varying pH (final [AP] = 155 pg/ml) and after 30- min aggregation was measured by sample Also. Plotted in the figure is the percentage of the total extent of aggregation of a given sample (measured by sample absorbance) with respect to what is seen at pH 6.5 where essentially 100% of the protein is aggregated. Also shown in the figure is a calculated titration curve (dashed line) for a weak acid with a pK. of 7.38, plotted as the percentage in the protonated form. The inset shows the extent of self-association as a function of A P concentration at pH 6.5 (0) and at pH 7.2 (A).

by guest on April 2, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4440 Self-Association of Assembly Protein AP-2

O i O ' I ' I I ' ' 1 0 20 30

Time (rnin) FIG. 5. Reversibility of AP aggregation. AP (1.16 mg/ml in

10 mM Tris-HC1, pH 8.5) was diluted 7.6-fold in 0.1 M sodium MES, pH 6.5 (0) or pH 7.2 (0) and the sample A350 was monitored as a function of time. The final AP concentration was 153 pg/ml. After 30 min the pH 6.5 sample was centrifuged (2 min at 10,000 X g) and the pellet was resuspended in the original volume of 10 mM Tris-HC1, pH 8.5. This sample was again diluted with 0.1 M sodium MES, pH 6.5 (0) or 7.2 (W) such that the final concentration of AP was again 153 pg/ml. The value at time zero represents the A350 obtained when an equivalent sample was appropriately diluted with 10 mM Tris, pH 8.5, and gives an indication of the extent of aggregation before dilution with MES.

1 2 S P S P

APlOOs [

AP50 - r

LM-AP

AP16 -

1 HM-AP

J FIG. 6. Identification of the AP-2 domain responsible for

self-association. Intact AP ( l a n e 1 ) and trypsin-treated AP ( l a n e 2) were diluted with buffer A (132 pg/ml final), incubated 30 min, centrifuged, and analyzed by gel electrophoresis. With proteolyzed AP-2 only the 16-, 50-, and 60-66-kDa polypeptides (designated heavy mero-AP (HM-AP)) self-associate, whereas the 30-40-kDa fragments (light mero-AP (LM-AP)) remain in the supernatant fraction (s).

determine which of these domains is responsible for the self- association interactions, proteolyzed AP was diluted into buffer A. A rapid increase in sample turbidity indicated that the proteolyzed sample retained its self-association activity (data not shown). On centrifugation, the heavy mero-AP pelleted while the light mero-AP remained exclusively in the supernatant (Fig. 6), indicating that the site or sites on AP-2 responsible for self-association are located within its large central domain.

DISCUSSION

We have found that molecules of AP-2 can undergo an extensive self-association reaction that can be detected quan- titatively by measuring sample absorbance due to turbidity. While it is unlikely that the resultant large aggregates formed in uitro are themselves biologically relevant structures, the AP-2 self-association process is consistent with the existence of discrete, specific, and high affinity interactions. Self-asso- ciation occurs rapidly, on relatively mild changes in pH and buffer conditions, and is readily and rapidly reversible. It is not driven by exposure of buried hydrophobic domains such as might accompany gross denaturation, as nonionic deter-

gents have no discernible effect on the self-association proc- ess. The specificity of the interactions are further supported by the observations that of the three structurally and func- tionally related bovine brain assembly proteins, only the AP- 2 undergoes self-association (Fig. 5).

Several of our observations indicate that AP-2 self-associ- ation may occur within the intact cell. The reaction occurs in uitro under conditions that approximate cytoplasmic pH; its pK, (7.4) suggests that self-association could be effectively modulated by small shifts in intracellular pH. Furthermore, the dissociation constant that is inferred from the concentra- tion dependence of self-association M) reflects a high affinity interaction. This affinity is equal or greater than that of other protein-protein interactions of physiological signifi- cance such as tubulin (33) or actin (36) polymerization, or the interactions of profilin with actin (24) or red blood cell band 4.1 with glycophorin (25). Although direct quantitation of AP- 2 in cells or tissues has not been reported, by comparison with known clathrin levels (26) the whole cell concentration of AP-2 in brain can be estimated to be approximately 30-60 pg/ml or 80-160 nM.4 These concentrations are well within the range in which AP-2 self-association has been observed in the experiments reported here. Moreover, cellular AP-2 has been localized to plasma membrane (14,20) and endosome surfaces (31). In this context, local concentration effects due to membrane binding have been shown to greatly amplify protein polymerization events, potentially by orders of mag- nitude (27), suggesting that the concentrations required to observe self-association in solution may represent an extreme upper limit.

Under the assay conditions used here, AP-1 and AP-3 are capable of assembling clathrin into coat structures but do not exhibit self-association. With respect to AP-2 in particular, in the accompanying paper (35) we show that molecules containing polyphosphate groups block self-association at concentrations that do not significantly affect AP-2-mediated coat assembly in solution. These findings indicate that self- association is not a general requirement for the in uitro assembly of clathrin lattices in solution. However, to the extent that the mechanism of lattice assembly on membrane surfaces may differ fundamentally from the solution process, it may be premature to completely rule out a role for the AP- AP interactions in the assembly of clathrin coat structures within the intact cell.

An alternative view is that self-association is a function of AP-2 independent of clathrin assembly. The observation that AP-2 can bind the cytoplasmic domain of several transmem- brane receptors (29, 30) suggests that self-association may contribute to receptor clustering at the cell surface, giving rise to the aggregated receptor intermediates seen early in the endocytic pathway. AP-2-induced receptor clustering might also occur on the surface of endosomes, where AP-2 has recently been localized (31), perhaps driving the segregation of receptors from released ligand seen within the tubulovesic- ular regions (32). These possibilities suggest that the novel self-association reaction described here may play important roles in AP-2 functions within cells.

Acknowledgments-We wish to thank J. Murphy, I. T. Pleasure, and K. Prasad for helpful discussions.

REFERENCES

1. Goldstein, J. L., Brown, M. S., Anderson, R. G. W., Russel, D. W., and Schneider, W. J. (1985) Annu. Reu. Cell Biol. 1, l -19

This assumes that the C1athrin:AP ratio in cells is similar to that present in isolated coated vesicles (2:1, Ref. 11).

by guest on April 2, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Self-Association of Assembly Protein AP-2 4441 2. Pastan. I. H.. and Willingham. M. C. (1985) in Endocytosis 19. Kohtz, D. S., and Puszkin, S. (1988) J. Biol. Chem. 263, 7418-

3.

4.

5. 6. 7.

8. 9.

10. 11. 12.

13.

14. 15.

16.

17.

18.

- . . . (Pastan, I. H., and Willingham, M. C., eds) pp. 1-40, Plenum Publishing Corp., New York

Pearse, B. M. F., and Crowther, R. A. (1987) Annu. Rev. Biophys. Biophys. Chem. 16,49-68

Dahms, N. M., Lobel, P., and Kornfeld, S. (1989) J. Biol. Chem.

Ungewickell, E., and Branton, D. (1981) Nature 289 , 420-422 Kirchhausen, T., and Harrison, S. C. (1981) Cell 23, 755-761 Keen, J. H., Willingham, M. C., and Pastan, I. H. (1979) Cell 16,

Zaremba, S., and Keen, J. H. (1983) J. Cell Biol. 97,1339-1347 Pearse, B. M. F., and Robinson, M. S. (1984) EMBO J. 3,1951-

Ahle, S., and Ungewickell, E. (1986) EMBO J. 5, 3143-3149 Keen, J. H. (1987) J. Cell Biol. 105, 1989-1998 Prasad, K., and Lippoldt, R. E. (1988) Biochemistry 27, 6098-

Ahle, S., Mann, A., Eichelsbacher, U., and Ungewickell, E. (1988)

Robinson, M. S. (1987) J. Cell Biol. 104,887-895 Keen, J. H. (1985) in Endocytosis (Pastan, I. H., and Willingham,

M. C., eds) pp. 85-129, Plenum Publishing Corp., New York Wiedenmann, B., Lawley, K., Grund, C., and Branton, D. (1985)

J. Cell Biol. 101, 12-18 Woodward, M. P., and Roth, T. F. (1978) Proc. Natl. Acad. Sci.

Keen, J. H., and Black, M. M. (1986) J. Cell Biol. 102, 1325-

264,12115-12118

303-312

1957

6104

EMBO J. 7,919-929

U. S. A. 75,4394-4398

1333

20.

21. 22. 23.

24.

25. 26.

27.

28. 29. 30.

31.

32.

33.

34. 35.

36.

7425 Bar-Zvi, D., Mosley, S. T., and Branton, D. (1988) J. Biol. Chern.

Heuser, J. E., and Keen, J. H. (1988) J. Cell Biol. 107,877-886 Zaremba, S., and Keen, J. H. (1985) J. Cell Biochern. 28,47-58 Keen, J. H., and Beck, K. A. (1989) Biochern. Biophys. Res.

Cornrnun. 158, 17-23 Stossel, T. P., Chaponnier, C., Ezzell, R. M., Hartwig, J. H.,

Janmey, P. A,, Kwiatowski, D. J., Lind, S. E., Smith, D. B.,

Rev. Cell Biol. 1 , 353-402 Southwick, F. S., Yin, H. L., and Zaner, K. S. (1985) Annu.

Anderson, R. A., and Lourien, R. F. (1984) Nature 307,655-658 Goud, B., Huet, C., and Louvard, D. (1985) J. Cell Biol. 100,

Grasberger, B., Minton, A. P., DeLisi, C., and Metzger, H. (1986)

Madshus, I. H. (1988) Biochem. J. 250,l-8 Pearse, B. M. F. (1988) EMBO J. 7, 3331-3336 Glickman, J. N., Conibear, E., and Pearse, B. M. F. (1989) EMBO

Guagliardi, L. E., Koppelman, B., Blum, J. S., Marks, M. S., Cresswell, P., and Brodsky, F. M. (1990) Nature 343, 133-139

Geuze, H. J., Slot, J. W., and Schwartz, A. L. (1987) J. Cell Biol.

263,4408-4415

521-527

Biochemistry 83,6258-6262

J. 8 , 1041-1047

104.1715-1723 Carlier, M. F., Hill, T. L., and Chen, Y. D. (1984) Proc. Natl.

Acad. Sei. U. S. A. 81. 771-775 Weisenberg, R. C. (1986) Ann. N.Y. Acad. Sci. 466, 543-551 Beck, K. A., and Keen, J. H. (1991) J. Biol. Chern. 266, 4442-

Wegner, A., and Isenberg, G. (1983) Proc. Natl. Acad. Sci U. S. 4447

A. 80,4922-4925

by guest on April 2, 2018

http://ww

w.jbc.org/

Dow

nloaded from

K A Beck and J H KeenAP-2.

Self-association of the plasma membrane-associated clathrin assembly protein

1991, 266:4437-4441.J. Biol. Chem. 

  http://www.jbc.org/content/266/7/4437Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/266/7/4437.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on April 2, 2018

http://ww

w.jbc.org/

Dow

nloaded from